Indian enterprises are rapidly advancing in artificial intelligence adoption and a growing cohort has embedded AI into the core of operations. Functions such as customer service, software development, analytics, product engineering and innovation are already delivering measurable returns, according to industry leaders at The Economic Times AI Vantage Roundtable on AI Infrastructure and Data Architecture in Bengaluru.
"We are seeing software development life cycles transforming. So from that perspective, we are certainly looking at faster, better, cheaper," said Sandhya Arun, chief technology officer, Wipro. However, the panelists said the broader business landscape is still adjusting to the AI-led transformation and workplace changes. "Technology has always been viewed as a tool for humans. And today, suddenly, we are looking at technology as a colleague. And that certainly requires a certain amount of mindset shift across the board," she said.
"The skill set required by the knowledge worker is fundamentally changing," Balaji Thiagarajan, chief technology and product officer, Flipkart said, adding organisations must address staff concerns on the future of work and workforce displacement.
Data infrastructure gaps
The panelists viewed that deeper AI integration faces challenges due to legacy data infrastructure, unequal access to computing resources and sustainability concerns. Daisy Chittilapilly, president, Cisco India and South Asia, said most organisations are attempting to build a "21st-century AI economy on a 20th-century infrastructure," creating a widening gap between AI ambitions and execution.
"Technical debt has become a big blocker," Chittilapilly said, adding that AI adoption is increasingly becoming a capital-intensive conversation centred on infrastructure investments, equitable access to compute resources and energy sustainability.
She said that while productivity benefits are increasingly visible, revenue-generation use cases remain less mature and may require enterprises to fundamentally rethink business models, workflows and service delivery.
Amiteshwar Seth, senior vice- president and global delivery head of AI and data, Cognizant, said AI merely amplifies existing organisational capabilities and weaknesses. "If your data is messy, what's going to come out after AI amplifies it is a bigger mess," Seth said, identifying data readiness, organisational culture, governance and strategic prioritisation as the primary barriers to successful AI adoption.
He cited industry estimates suggesting that only 6-8% of AI projects ultimately achieve their intended business outcome, with most initiatives failing to move beyond proof-of-concept stages due to poorly defined objectives, integration challenges and underestimation of operational costs.
Business model transformation
Mritunjay Singh, chief operating officer, L&T Technology Services, said the current AI wave was fundamentally different from earlier technology transitions because it had the potential to reshape the economics of businesses.
"If you look at the AI evolution, the new wave is far more disruptive. It is going to reimagine the work and the way it is being done," Singh said. He said once AI begins restructuring task execution, it will have a direct impact on companies' cost structures, revenue models and profitability.
Lakshminarayanan Ramalingam, chief operating officer, Quest Global, argued that AI adoption should ultimately be measured not by pilot projects but by whether companies can materially alter business models, reduce cost structures and create new revenue streams. LTTS' Singh echoed that view.
"The way to look at it is how many companies are able to change cost structure by 40%, revenue expansion by 40% or 60%. Unless the disruption reaches a business model, you're not really talking about adoption. We are talking about experimenting," Singh said.
India's opportunity
Cisco's Chittilapilly argued that India's greatest opportunity lies not necessarily in building the largest frontier AI models but in developing domain-specific applications and smaller language models tailored to industry needs. "The value will eventually shift from infrastructure layers to applications," she said. "That is where India's technology industry can make a significant contribution."
Thiagarajan echoed that view, calling AI a once-in-a-generation opportunity similar to India's leapfrog from limited fixed-line telephony to widespread mobile connectivity.
"This is our opportunity to take or lose," he said.
Kishore Alva, president and executive director of Adani Group in Karnataka, said AI infrastructure should be viewed as a long-term strategic investment rather than a technology expenditure. He highlighted the importance of power availability, renewable energy, water resources and land availability in supporting future AI infrastructure growth.
Quest's Ramalingam also said that India should develop language models.
"We should collaborate with everybody. We should keep our market open so that the best minds come and experiment...We should get our manufacturing technology for chip making. We should get our own eventual LLMs."
"It is easier to say it is expensive, but at some point in time, it is going to be not so differentiated. We should do those, so we do not end up paying royalty elsewhere. And these are things that we should do step by step."
(This article is part of the AI Vantage series, developed in partnership with Cisco)
